This paper presents a method that uses the threaded dental implant as a reference object for the inter-image alignment necessary for digital subtraction radiography. The implant is furthermore used to define a measurement coordinate system and to automate the placement of reference areas used for contrast correction. The method is intended for studies of diffuse bone density changes in the vicinity of the implant. The method is shown to be insensitive to large variations in exposure time and geometry, and is together with the contrast correction method of Ruttimann et al., able to detect clinically invisible simulated bone density changes.
A common question in movement studies is how the results should be interpreted with respect to systematic and random errors. In this study, simulations are made in order to see how a rigid body's orientation in space (i.e. helical angle between two orientations) is affected by (1) a systematic error added to a single marker (2) a combination of this systematic error and Gaussian white noise. The orientation was estimated after adding a systematic error to one marker within the rigid body. This procedure was repeated with Gaussian noise added to each marker.
In conclusion, results show that the systematic error's effect on estimated orientation depends on number of markers in the rigid body and also on which direction the systematic error is added. The systematic error has no effect if the error is added along the radial axis (i.e. the line connecting centre of mass and the affected marker).